When the Carmignano Visitation was carried over the threshold of my studio, Pontormo was certainly not a new visitor. In 2010, on the occasion of the Bronzino exhibition held at the Palazzo Strozzi,1 I restored three of the four tench in the pen-dentives of the Capponi chapel on-site in the church of Santa Felicita in Florence. Those procedures were different and special because each work of art requires (and deserves!) a customized preliminary study before proceeding with the restoration proper. Therefore, the plan of action begins with something the visitor usually does not see—the support (fig. 1) and then proceeds layer by layer, from the ground to the colors to the varnishes that cover it all. This approach, which is not necessarily universal, will be the leitmotif of this essay, allowing us to feel the embrace and walk along the multicolored paths of maestro Jacopo.
Wood, a living material, deserves its own story, which begins with the most widely used species for Florentine paintings: poplar. When the "skin" fell off the stacked and seasoned planks, they were cut to shape. Five planks, the straightest (four quarter cut and one slab cut), were shaped using two tools: they were rough-cut with an axe and then finely whittled with a "curved, clean-cutting blade."The planks are 206 cm long and vary in width from the narrowest, which is 16 cm, to the widest, 38 cm. From this standpoint, no two are identical.
Initially the wood was a light blond color, but the years have turned it nearly brown. There are still coarse bristles in some areas (fig. 2), and two of the boards still present the slabwood (sciavero), the part just beneath the bark that was left from cutting the log. The knots in the grain were not removed, even though it was well known that they could damage the surface over the long term. Other half-empty knots were removed and the holes were filled with dia-mond-shaped poplar inserts that are still discernible under the paint layer.
Then, using a narrow-blade planer, three, approximately 1.5 cm parallel "scoring lines" were cut 40 cm apart (fig. 3); they served as a guide to ensure that the planks would be planed to the same thickness, and the panel would be as we see it today.
The planks were attached with cold casein glue, nine rectangular hardwood inserts called ranghette set into the heart of each plank, and two cylindrical dowels fitted into each. The small heads of the dowels still peek out on the painted side. Lastly, the two larch dovetail battens were inserted near the upper and lower edges of the panel; they are tapered in the same direction, narrower at the insertion end and wider at the tail. Initially they were supposed to be identical but, prior to insertion, a 1 cm strip was added to the lower one for better support. There are fifty-three butterfly inserts. These were set into the joins, most probably during a restoration conducted prior to the 1956 exhibition. The final phase was the priming with several layers of viscous white gesso and hot glue; these typically drip onto the outer edges during their application and their presence confirms that the panel's dimensions have not been altered.
Jacopo's preliminary ideas have fortunately survived in a small but exquisite drawing (32.6 cm long by 24 cm wide, cat. 4) conserved in the Gabinetto dei Disegni e delle Stampe degli Uffizi (inv. 461 F). The drawing shows the entire composition with a wealth of details, done in black chalk with slightly sfumato lines, with the more strongly shaded areas defined by parallel, oblique hatching. The squaring on the front of the sheet was done in red chalk, and consists of fourteen vertical and ten
horizontal lines that define 20.9 by 20.3 mm squares. Following a common practice of the time, the drawing was transferred by drawing freehand in charcoal on the well:smoothed gesso following the "grid,"2 which was enlarged seven and a half times to match the final dimensions of the altar niche for which it was presumably intended. The degree of detail in the drawing varies according to the importance of the subject. The architecture, for example, is more schematic: the building on the right with the balustrade is sketched with only seven vertical lines, which were then incised directly onto the panel, while the steps are formed by the same number of horizontal lines. Since the left side of the sheet has been trimmed, all we see of the building is a cubic portion and the tower, and we have no way of knowing how faithful the composition that Pontormo transferred to the primed board or that is visible in the final painting is to the original drawing. It may have evolved while the artist worked on the panel.
We can compare the drawing on paper and the final version of the painting using infrared reflectography,3 which allows us to penetrate the paint surface and compare the drawing to how it was first transferred to the gesso ground (fig. 4). This analysis revealed a series of the artist's pentimenti and second thoughts, allowing us to imagine Pontormo while he pondered and reconsidered the composition, deciding which de-tails to incise: barely sketched eyes like black dots, mouths that seem to be perpetually opening and closing, hands with softly clenched fingers, feet that shift in the search for new and unstable movements, like those of the attendant on the left, which were rethought over and over. We will start with the young attendant (fig. 5). In the drawing, her face and hair are tilted more to the left and her face is rounder. The eyes, turned in the same direction and not toward the viewer were executed by drawing a simple circle and then adding the pupils. The veil seems to be carelessly twisted, leaving her hair free to be tousled by the wind. This is a clear contrast to the marked composure of the same figure in both the underdrawing and the painting. Her robe is almost identical in both versions except that in the painting it adheres much more closely to her thigh. The feet, as already noted, show the most pentimenti. In the drawing they are seen frontally, overlapping and slightly shifted; the reflectogram, on the other hand, shows a foot that was first turned three-quarters, and then repositioned in a more frontal solution (fig. 6).
The old attendant's face is broader, sadder, and less oval in the drawing than on the panel, where the veil covers her head to a greater extent. Here, the reflectogram shows a foot drawn with a few decisive lines, whereas in the drawing the entire lower part is a confused tangle of drapery. In the drawing, the Virgin Mary's blue mantle is simply folded at the shoulder over the sleeve, while on the panel the veil comes down her neck, covering her ear to rest on the shoulder. The hand on her cousin's shoulder is less tapered even though the opening of the index finger is the same. The legs, seen in profile, are more bent and dynamic, the cloak adhering more closely to the left calf, forming a V-shaped crease, which on the panel was transformed into two folds and a tangle of drapery that falls down over the heel of the right foot. That foot was reworked three times on the draw-ing and twice on the panel, and the pink sleeve was reduced and covered by the blue cloak down to the middle of the forearm. The least-changed figure is Saint Elizabeth. In all the versions, the red ocher robe and the head covering are almost fold-by-fold identical, as is the way the fabric rolls at her thigh. In the drawing her face and neck stretch closer to the Virgin, and her mouth has a more serious expression. The left foot shows some changes in its angle, which, however, are not visible in the reflectogram. The right foot is parallel to the ground and covered obliquely by the orange mantle that drags on the street, whereas in the painting the hem has been shortened to reveal the older attendant's foot.
In both the modello and the underdrawing, the hand with which she clutches Mary's waist has the thumb raised, while the painting reveals one of Pontormo's pentimenti: the thumb was repainted over the blue and rests in the shaded fold beneath the armpit (fig. 7). The three-dimensional volumes of the buildings and pitched roofs are more impos-ing than in the drawing and were incised in the ground layer, while the small genre scenes- the two slender male figures near the entrance to the house and the woman at the window above them—were added later. The doorjamb, the balustrade, and the stairs were all directly incised into the dry gesso with a fine point.
The time had come for Pontormo to color his figures. It is reasonable to assume that the likely patron, Bonaccorso Pinadori, who according to Vasari had his portrait painted by Bronzino,4 supplied the artist with pigments and paintbrushes, since that was his business. The pigments, whether pure or mixed, were combined with an oil medium 5 and applied with bold brushstrokes in overlapping, at times almost transparent layers. It is not easy to establish the precise order in which Pontormo applied his colors: we can only hypothesize about his general approach, thanks mainly to observations under raking light of the partial and random overlap of the impasto. It is likely that after incising the architectural lines in the gesso ground with a fine stylus, Pontormo began blocking out the buildings and rocky ground with brushstrokes of lead white mixed with carbon black, adjusting the gray tones to simulate pietra serena and adding yellow ocher for the pietra forte, including the prism-shaped structure (perhaps the demolished prison known as the Carcere delle Stinche, fig. 8) behind Saint Elizabeth, the first figure he painted. He began with her orange cloak,6 painting straight off and softening the folds in the light with tin yellow and lead white,7 and the half-tones and shadows by gradually adding brown earths and cinnabar. He then moved on to the rest, painting the Virgin's hand on the shoulder and the ash gray8 veil with continuously alternating layers of color, ending with the quick brushstroke for the thumb, then tightening the sash and filling the negative spaces of her costume with malachite green,9 ending with the flesh tones of the face and foot. The faces of all four women (fig. 9) took shape quickly with thick, freehand strokes of lead white and cinnabar using a flat brush. Fine, fluid lines outline the pupils (always consisting of a dark dot) and strong contour lines delineate the eyes. A few touches of lead white and green earth imbue the eyes with a fixed yet languid stare, while the noses and mouths are outlined with cinnabar, producing the effect of thin flame-like strokes of red light from below.
A fine, light gray line that is actually the color of the old attendant's veil high-lights Mary's profile Jacopo's original invention to increase the relief of his figures. Still using oily colors, Jacopo gave substance to what can almost be considered "an appendix" to the figure of the Virgin, a double but viewed from the front: the older attendant. She has the same veil, the same face, but less brightly lit, and painted with the same technique. Over the following days, and we do not know how many, he began painting the other two figures, the Virgin Mary and the young attendant. Once again, he mixed the flesh tones using lead white and cinnabar, and painted Mary's face and a light layer of hair. He carefully dressed her in that soft cyclamen veil, woven with white lead and rose madder,10 adding a jaunty glimpse of pink sleeve. Next came the billowing azurite blue cloak11 (fig.10).
He painted swiftly, and the brush wrinkled the folds, the deepest shadows overlapping parallel and crosshatched lines. He sometimes used the butt of the brush to scratch the paint surface, softening the fabric with small. rounded grooves. The young attendant's face was designed following the model of her older companion: first a breath of life, then the pink mantle, whose hem reaches down to the Virgin's foot, the green lacquer turban that spills onto the buildings, what may be a hat held to her chest, and lastly the vermilion hair. But Jacopo was still not satisfied: he stepped back, looked, and decided that the figure was impossible. He had to decide which of the two underlying figures to use, and in the end he raised the mouth, eyes, and nose, opted for a less voluminous headdress, and covered the frontally posi-tioned foot with a few more centimeters of robe. Next came the sky,12 which he had already sketched with diluted brushstrokes, now silhouetting the faces and buildings, including the little triangle beneath the Virgin's chin.
He completed the painting with the pavement, ground, and some last details: the attendant's blond curls fluttering in the wind, the two men sitting on the bench, the female figure at the window13 (fig. 11), the donkey's head peeking out from behind the corner, and finally the two halos floating above the cousins' heads.
The first step in the restoration was a visual inspection of the painted surface14The restorer's eye is the first instrument for recognizing color changes, cracks, and scratches that document the events and restorations the painting underwent over the years. The many studies conducted on the painting — via Osiris camera infrared reflectography,15 false color infrared photography (see fig. 5), ultraviolet fluorescence (see fig. 5 again), XRF (X-ray fluorescence), examining samples under optical and SEM (scanning electron microscope) with a microprobe, EDS (energy-dispersive X-ray spectroscopy), and GC-MS gas chromatography specific for the media — provided further invaluable information about Pontormo's techniques and the composition of the materials he used. Although the panel has maintained its original shape and size, some inserts had been set in between the joins, into spaces slightly larger than the butterfly inserts. These had not compromised the equilibrium between the original panel and the inserts themselves, and therefore did not have to be removed. The micro-cracks be-tween the planks were repaired by inserting thin wedges (some even fitted between the butterfly inserts) of aged poplar, which were then secured at the ends of the panel with vinyl glue. The cleaning was done with emulsions of bleached wax and solvents,16 carefully gauged to thin rather than totally remove the varnishes that probably include a sandarac resin embedded into the blue of the Virgin's cloak, which is thus still slightly whitish in tone. Some mechanical damage caused by moving the painting, such as nicks and scratches, had been filled with carnauba wax during previous restorations, as were the 1,673 wormholes that were emptied individually and refilled (fig. 12).17 The overpainting in oils that hid some of the original details was removed after a complex cleaning done with the aid of an optical microscope (fig. 13).18 The most heavy-handed overpainting was in the sky above the heads of the figures, the buildings to either side, the ground below, and the shadows of the four women's feet. Inpainting was done with watercolors and temperas, while final retouching was executed after applying Retoucher protective varnish with a brush and atomizer.
The two tradesmen, who reemerged intact from beneath the overpainting of the bench, are once again engaged in their conversation, while above them, from the window of the once-abandoned palazzo, a Florentine housekeeper hangs a cloth as she did nearly five hundred years ago. But the most delightful and unexpected figure to reemerge is the one peeking out from behind the corner of the same building: a curi-ous little gray donkey—a living form returned to a metaphysical background—that master Jacopo included to add a smile to the sacred encounter (fig. 14).
I would like to thank the friends and colleagues who participated in this complex restoration at different times and in different ways. In particular Gloria Verniani, my trusted assistant; Cristina Gnoni Mavarelli, who directed the project; Fabrizio Moretti, who financed the restoration; Helena Bernal, Luigina Ciurla, Elena Cupisti, and Umi Toyosaki, trusted helpers; Roberto Buda for his work and suggestions on the wooden support; Teobaldo Pasquali for the diagnostic studies; Alfredo Aldrovandi for being so generous with his time; Azzurra Macherelli and Francesca Briani of ADARTE; Marco Zanaboni and Francesca Modugno of the Department of Chemistry and Industrial Chemistry of the University of Pisa for the chemical analyses; Antonio Quattrone for the photographic documentation; Giulio Aspettati for the videos; Luca Mattedi and Davide Civettini for their constant contributions; and Bruce Edelstein, Davide Gasparotto. and Michele Grasso for their valuable advice and suggestions.
1 A. Baldinotti, in Bronzino 2010, 60-65 nos. 1.3, 1.4, 1.5, 1.6.
2 The drawing, or model, was squared, making it possible to transfer the drawing to a larger sheet or to the primed panel line by line, square by square, enlarging the composition while maintaining the original proportions of the drawing.
3 Infrared radiation has a wavelength ranging from 1 to 2 microns; the waves penetrate the paint layer and are reflected and captured by the camera. The black-and-white image is immediately visible on the monitor, The image makes it possible to see the preparatory drawings beneath the paint layer and identify the artist's pentimenti, as well as conservation condition, retouches, filler, and losses of color. However, these traces are only visible if the artist used a charcoal-based medium for the underlying drawing.
4 Vasari, ed. deVere, II, 870; see also Vasari 1568, ed. Bettarini-Barocchi 1966-97, VI, 1987, 232.
5 The tests, done on a specimen using the GC-MS supplied by ADARTE Firenze at the department of Chemistry and Industrial Chemistry of the University of Pisa, showed that he used linseed oil as the binder. The same section was analyzed under the SEM, which showed that the blue particles are azurite, the white in which they are immersed is lead white, and the red particles are red lacquer.
6 XRF spectrophotometry is a noninvasive technique that makes it possible to identify the elements in a specimen by studying its X-ray fluorescence. The radiation is emitted by the atoms after they have been excited (this may also create a photoelectric effect); this is done by irradiating the sample with high-energy gamma and X-rays. Similar effects can be obtained using ion beams. For the orange robe, XRF analysis: lead, with a smaller amount of mercury (relatable to cinnabar, mercury sulfide) and iron (indicative of a red or yellow ocher).
7 XRF analysis: lead (lead white), tin (lead yellow and tin).
8 XRF analysis: lead and iron, relatable to lead white and ocher.
9 XRF analysis: copper, therefore possibly malachite (natural, mineral basic copper carbonate), with a small amount of lead (probably lead white).
10 XRF analysis: lead, while the smaller amount of copper could be from the underlying robe.
11 XRF analysis: copper (with a smaller amount of lead), indicative of azurite.
12 XRF analysis: considerable amount of lead along with a small quantity of copper, equivalent to azurite mixed with lead white.
13 It is possible that the building where we see the woman leaning from the window is the same one that appears in the Nerli Altarpiece by Filippino Lippi in the church of Santo Spirito in Florence, or that it was inspired by that painting.
14 The restoration lasted from November 2013 to March 2014.
15 Osiris camera with InGaAs line sensor, 900 to 1700 nm sensitivity.
16 In this case we used solvents such as acetone, isopropyl alcohol, triethanolamine (TEA), and white spirit.
17 Filled with Polyfilla injected into the hole followed by gesso and rabbit skin glue.
18 We used a Nikon optical microscope with fiber optics and triethanolamine dispersed in a waxy emulsion, neutralized with ligroin.